ASCO: An Availability-aware Service Chain Orchestration

Network Functions (NFs) are decoupled from dedicated hardware equipment and present a better shareability of physical resources on basis of Network Function Virtualization (NFV) technology. The placement and routing of VNFs are orchestrated flexibly by operators for providing differentiated services to users. QoS parameters such as delay, service availability and resource utilization efficiency need to be taken into account. In this paper, an availability-aware service chain orchestration (ASCO) mechanism including pre-pruned depthfirst search (PDFS) algorithm and M-to-N backup mechanism is proposed. PDFS is developed from depth-first search algorithm by taking delay as pruning factor. An end-to-end reachable and cost-optimized path set is obtained and extended to final orchestration path through adopting of PDFS. Moreover, M-to-N backup mode allows one of M primary VNFs to have N migration options. Load balancing factor is designed to achieve network traffic balance by selecting path with more remaining bandwidth. The simulation result shows that our algorithm can meet QoS requirements and service availability in a low-cost way.